Exploration of Machine Learning Classification Models Used for Behavioral Biometrics Authentication

TL;DR

This paper reviews machine learning algorithms used in behavioral biometric authentication on mobile devices, focusing on touch dynamics and movement, highlighting benefits, limitations, and future research directions.

Contribution

It provides a comprehensive review of current ML algorithms in behavioral biometric mobile authentication, emphasizing touch and movement data.

Findings

Identifies key ML algorithms used in behavioral biometrics

Discusses benefits and limitations of current approaches

Recommends future research directions

Abstract

Mobile devices have been manufactured and enhanced at growing rates in the past decades. While this growth has significantly evolved the capability of these devices, their security has been falling behind. This contrast in development between capability and security of mobile devices is a significant problem with the sensitive information of the public at risk. Continuing the previous work in this field, this study identifies key Machine Learning algorithms currently being used for behavioral biometric mobile authentication schemes and aims to provide a comprehensive review of these algorithms when used with touch dynamics and phone movement. Throughout this paper the benefits, limitations, and recommendations for future work will be discussed.